language.go

     1  // Copyright 2013 The Go Authors. All rights reserved.
     2  // Use of this source code is governed by a BSD-style
     3  // license that can be found in the LICENSE file.
     4  
     5  //go:generate go run gen.go gen_index.go -output tables.go
     6  //go:generate go run gen_parents.go
     7  
     8  package compact
     9  
    10  // TODO: Remove above NOTE after:
    11  // - verifying that tables are dropped correctly (most notably matcher tables).
    12  
    13  import (
    14  	"strings"
    15  
    16  	"golang.org/x/text/internal/language"
    17  )
    18  
    19  // Tag represents a BCP 47 language tag. It is used to specify an instance of a
    20  // specific language or locale. All language tag values are guaranteed to be
    21  // well-formed.
    22  type Tag struct {
    23  	// NOTE: exported tags will become part of the public API.
    24  	language ID
    25  	locale   ID
    26  	full     fullTag // always a language.Tag for now.
    27  }
    28  
    29  const _und = 0
    30  
    31  type fullTag interface {
    32  	IsRoot() bool
    33  	Parent() language.Tag
    34  }
    35  
    36  // Make a compact Tag from a fully specified internal language Tag.
    37  func Make(t language.Tag) (tag Tag) {
    38  	if region := t.TypeForKey("rg"); len(region) == 6 && region[2:] == "zzzz" {
    39  		if r, err := language.ParseRegion(region[:2]); err == nil {
    40  			tFull := t
    41  			t, _ = t.SetTypeForKey("rg", "")
    42  			// TODO: should we not consider "va" for the language tag?
    43  			var exact1, exact2 bool
    44  			tag.language, exact1 = FromTag(t)
    45  			t.RegionID = r
    46  			tag.locale, exact2 = FromTag(t)
    47  			if !exact1 || !exact2 {
    48  				tag.full = tFull
    49  			}
    50  			return tag
    51  		}
    52  	}
    53  	lang, ok := FromTag(t)
    54  	tag.language = lang
    55  	tag.locale = lang
    56  	if !ok {
    57  		tag.full = t
    58  	}
    59  	return tag
    60  }
    61  
    62  // Tag returns an internal language Tag version of this tag.
    63  func (t Tag) Tag() language.Tag {
    64  	if t.full != nil {
    65  		return t.full.(language.Tag)
    66  	}
    67  	tag := t.language.Tag()
    68  	if t.language != t.locale {
    69  		loc := t.locale.Tag()
    70  		tag, _ = tag.SetTypeForKey("rg", strings.ToLower(loc.RegionID.String())+"zzzz")
    71  	}
    72  	return tag
    73  }
    74  
    75  // IsCompact reports whether this tag is fully defined in terms of ID.
    76  func (t *Tag) IsCompact() bool {
    77  	return t.full == nil
    78  }
    79  
    80  // MayHaveVariants reports whether a tag may have variants. If it returns false
    81  // it is guaranteed the tag does not have variants.
    82  func (t Tag) MayHaveVariants() bool {
    83  	return t.full != nil || int(t.language) >= len(coreTags)
    84  }
    85  
    86  // MayHaveExtensions reports whether a tag may have extensions. If it returns
    87  // false it is guaranteed the tag does not have them.
    88  func (t Tag) MayHaveExtensions() bool {
    89  	return t.full != nil ||
    90  		int(t.language) >= len(coreTags) ||
    91  		t.language != t.locale
    92  }
    93  
    94  // IsRoot returns true if t is equal to language "und".
    95  func (t Tag) IsRoot() bool {
    96  	if t.full != nil {
    97  		return t.full.IsRoot()
    98  	}
    99  	return t.language == _und
   100  }
   101  
   102  // Parent returns the CLDR parent of t. In CLDR, missing fields in data for a
   103  // specific language are substituted with fields from the parent language.
   104  // The parent for a language may change for newer versions of CLDR.
   105  func (t Tag) Parent() Tag {
   106  	if t.full != nil {
   107  		return Make(t.full.Parent())
   108  	}
   109  	if t.language != t.locale {
   110  		// Simulate stripping -u-rg-xxxxxx
   111  		return Tag{language: t.language, locale: t.language}
   112  	}
   113  	// TODO: use parent lookup table once cycle from internal package is
   114  	// removed. Probably by internalizing the table and declaring this fast
   115  	// enough.
   116  	// lang := compactID(internal.Parent(uint16(t.language)))
   117  	lang, _ := FromTag(t.language.Tag().Parent())
   118  	return Tag{language: lang, locale: lang}
   119  }
   120  
   121  // nextToken returns token t and the rest of the string.
   122  func nextToken(s string) (t, tail string) {
   123  	p := strings.Index(s[1:], "-")
   124  	if p == -1 {
   125  		return s[1:], ""
   126  	}
   127  	p++
   128  	return s[1:p], s[p:]
   129  }
   130  
   131  // LanguageID returns an index, where 0 <= index < NumCompactTags, for tags
   132  // for which data exists in the text repository.The index will change over time
   133  // and should not be stored in persistent storage. If t does not match a compact
   134  // index, exact will be false and the compact index will be returned for the
   135  // first match after repeatedly taking the Parent of t.
   136  func LanguageID(t Tag) (id ID, exact bool) {
   137  	return t.language, t.full == nil
   138  }
   139  
   140  // RegionalID returns the ID for the regional variant of this tag. This index is
   141  // used to indicate region-specific overrides, such as default currency, default
   142  // calendar and week data, default time cycle, and default measurement system
   143  // and unit preferences.
   144  //
   145  // For instance, the tag en-GB-u-rg-uszzzz specifies British English with US
   146  // settings for currency, number formatting, etc. The CompactIndex for this tag
   147  // will be that for en-GB, while the RegionalID will be the one corresponding to
   148  // en-US.
   149  func RegionalID(t Tag) (id ID, exact bool) {
   150  	return t.locale, t.full == nil
   151  }
   152  
   153  // LanguageTag returns t stripped of regional variant indicators.
   154  //
   155  // At the moment this means it is stripped of a regional and variant subtag "rg"
   156  // and "va" in the "u" extension.
   157  func (t Tag) LanguageTag() Tag {
   158  	if t.full == nil {
   159  		return Tag{language: t.language, locale: t.language}
   160  	}
   161  	tt := t.Tag()
   162  	tt.SetTypeForKey("rg", "")
   163  	tt.SetTypeForKey("va", "")
   164  	return Make(tt)
   165  }
   166  
   167  // RegionalTag returns the regional variant of the tag.
   168  //
   169  // At the moment this means that the region is set from the regional subtag
   170  // "rg" in the "u" extension.
   171  func (t Tag) RegionalTag() Tag {
   172  	rt := Tag{language: t.locale, locale: t.locale}
   173  	if t.full == nil {
   174  		return rt
   175  	}
   176  	b := language.Builder{}
   177  	tag := t.Tag()
   178  	// tag, _ = tag.SetTypeForKey("rg", "")
   179  	b.SetTag(t.locale.Tag())
   180  	if v := tag.Variants(); v != "" {
   181  		for _, v := range strings.Split(v, "-") {
   182  			b.AddVariant(v)
   183  		}
   184  	}
   185  	for _, e := range tag.Extensions() {
   186  		b.AddExt(e)
   187  	}
   188  	return t
   189  }
   190  
   191  // FromTag reports closest matching ID for an internal language Tag.
   192  func FromTag(t language.Tag) (id ID, exact bool) {
   193  	// TODO: perhaps give more frequent tags a lower index.
   194  	// TODO: we could make the indexes stable. This will excluded some
   195  	//       possibilities for optimization, so don't do this quite yet.
   196  	exact = true
   197  
   198  	b, s, r := t.Raw()
   199  	if t.HasString() {
   200  		if t.IsPrivateUse() {
   201  			// We have no entries for user-defined tags.
   202  			return 0, false
   203  		}
   204  		hasExtra := false
   205  		if t.HasVariants() {
   206  			if t.HasExtensions() {
   207  				build := language.Builder{}
   208  				build.SetTag(language.Tag{LangID: b, ScriptID: s, RegionID: r})
   209  				build.AddVariant(t.Variants())
   210  				exact = false
   211  				t = build.Make()
   212  			}
   213  			hasExtra = true
   214  		} else if _, ok := t.Extension('u'); ok {
   215  			// TODO: va may mean something else. Consider not considering it.
   216  			// Strip all but the 'va' entry.
   217  			old := t
   218  			variant := t.TypeForKey("va")
   219  			t = language.Tag{LangID: b, ScriptID: s, RegionID: r}
   220  			if variant != "" {
   221  				t, _ = t.SetTypeForKey("va", variant)
   222  				hasExtra = true
   223  			}
   224  			exact = old == t
   225  		} else {
   226  			exact = false
   227  		}
   228  		if hasExtra {
   229  			// We have some variants.
   230  			for i, s := range specialTags {
   231  				if s == t {
   232  					return ID(i + len(coreTags)), exact
   233  				}
   234  			}
   235  			exact = false
   236  		}
   237  	}
   238  	if x, ok := getCoreIndex(t); ok {
   239  		return x, exact
   240  	}
   241  	exact = false
   242  	if r != 0 && s == 0 {
   243  		// Deal with cases where an extra script is inserted for the region.
   244  		t, _ := t.Maximize()
   245  		if x, ok := getCoreIndex(t); ok {
   246  			return x, exact
   247  		}
   248  	}
   249  	for t = t.Parent(); t != root; t = t.Parent() {
   250  		// No variants specified: just compare core components.
   251  		// The key has the form lllssrrr, where l, s, and r are nibbles for
   252  		// respectively the langID, scriptID, and regionID.
   253  		if x, ok := getCoreIndex(t); ok {
   254  			return x, exact
   255  		}
   256  	}
   257  	return 0, exact
   258  }
   259  
   260  var root = language.Tag{}
   261
View as plain text